nasa vision 2030 cfd code · 2018-01-24 · nasa contracting officer representative – bil kleb...

Juan Alonso Stanford University

David Darmofal Massachusetts Inst. Of

Technology

William Gropp National Center for

Supercomputing Applications

Elizabeth Lurie Pratt & Whitney – United

Technologies

Dimitri Mavriplis University of Wyoming

Jeffrey Slotnick Principal Investigator Boeing Research & Technology [email protected]

Abdi Khodadoust Project Manager Boeing Research & Technology [email protected]

NASA Vision 2030 CFD Code Final Technical Review

Contract # NNL08AA16B (Order # NNL12AD05T)

Deliverable # 6

November 14, 2013 NASA Langley Research Center

mailto:[email protected]

mailto:[email protected]

2

Outline

Acknowledgments

Introductions

Overview of Study

Findings

Vision

Technology Development Plan

Recommendations

Answers to Key NASA questions

3

Acknowledgments

Extended Vision 2030 Team:

Joerg Gablonsky, Mori Mani, Robert Narducci, Philippe Spalart, and Venkat Venkatakrishnan – The Boeing Company

Robert Bush – Pratt & Whitney

NASA Technical Monitor – Mujeeb Malik

NASA Contracting Officer Representative – Bil Kleb

All attendees at the Vision 2030 CFD Workshop (May 2013),

especially our invited speakers:

Paul Durbin – Iowa State University

Sharath Girimaji – Texas A&M University

Brian Smith – Lockheed Martin Corporation

All who participated in the Vision 2030 CFD Survey

4

Introductions

5

Overview of Study

“...Address(es) the long range planning required by NASA’s Revolutionary Computational Aerosciences (RCA) sub-project, (which is managed under the Aeronautical Sciences Project) of the Fundamental Aeronautics Program (FAP)”

“…provide a knowledge-based forecast of the future computational capabilities required for turbulent, transitional, and reacting flow simulations…”

“…and to lay the foundation for the development of a future framework/environment where physics-based, accurate predictions of complex turbulent flows, including flow separation, can be accomplished routinely and efficiently in cooperation with other physics-based simulations to enable multi-physics analysis and design.”

6

Overview of Study CONTINUED

Create a comprehensive and enduring vision of CFD technology and capabilities:

Define/develop CFD requirements

Identify shortcomings and impediments

Develop a long-term, actionable research plan

Develop a detailed technology development roadmap to

– capture anticipated technology trends and future technological challenges,

– guide investments for long-term research activities,

– and provide focus to the broader CFD community for future research activities

“… the roadmap will aid and support NASA’s long-range research planning of FAP program elements and exploit developments in physical modeling, numerical methods, software, and computer hardware to develop a system-level view of the technology required for a 2030 CFD code”

7


Develop and execute a comprehensive CFD community survey to refine the technical requirements, gaps, and impediments

Based on the refined vision, hold a CFD workshop among subject matter experts within industry, government, and academia to: Get a definitive view on the relative priorities and importance of the

impediments that need to be overcome

Develop realistic options for technical approaches and ideas for improving the CFD capability

Assemble suggested plans for maturing CFD technologies through the Technology Readiness Level (TRL) scale, including requirements for validation and computational resources

Develop a preliminary timeline for research

Develop and deliver a final report summarizing findings and recommendations

8


Key NASA Questions:

1. What hardware requirements and software attributes will characterize an advanced “Vision 2030” CFD code that is used to routinely compute complex turbulent flows, involving flow separation, at subsonic, supersonic, and hypersonic speeds, and exhibits greater robustness than current technology?

2. How many orders of magnitude faster (time-to-solution), as compared with current capabilities, will 2030 CFD technology be?

3. What will be the fundamental elements of 2030 turbulence models, and for what classes of problems will they be reliable?

4. What are the principal impediments, both modeling and algorithmic, that must be overcome to achieve the 2030 CFD vision?

5. What high-risk/high-yield obstacles remain as enduring and daunting challenges that should be the focus of long-range efforts by NASA?

Findings

1. NASA investment in basic research and technology development for simulation-based analysis and design has declined significantly in the last decade and must be reinvigorated if substantial advances in simulation capability are to be achieved.

Physics-based simulation is a cross-cutting technology that impacts all of NASA aeronautics missions and vehicle classes – NAE Decadal Survey

Advances in simulation capabilities are often driven by the requirement of short-term impact, or in response to simulation failure on a program results in incremental improvements to CFD software

International government agencies (e.g., DLR, ONERA) routinely have base research and technology (R/T) elements as part of their long-term technology strategies

NASA’s Revolutionary Computational Aerosciences (RCA) project is a step in the right direction and should be maintained and expanded

9

Findings CONTINUED

2. HPC hardware is progressing rapidly and technologies that will prevail are difficult to predict. Current advances in exascale hardware architecture involve scalar

processors with 1000s of “streaming” processor cores, highly parallel memory interfaces, and advanced interconnects focus is on power consumption and failure recovery

Advanced software programming environments with higher levels of software abstraction will be required

Current CFD tools and processes do not scale well on these systems improved software development, implementation, and testing is needed

Alternative computing architectures are under development: – Quantum computers. Much focused attention, but application to CFD

is decades away.

– Superconducting logic. Some initial demonstration. Low densities, high costs.

– Low-power memory. Novel designs (e.g., cryogenics). Potentially higher latencies and densities.

– Massively parallel molecular computing. Initial demonstrations (e.g., nanosecond biological simulation of cell synthesis). Promises speeds similar to quantum computers.

10

Findings CONTINUED

3. The accuracy of CFD in the aerospace design process is severely limited by the inability to reliably predict turbulent flows with significant regions of separation

No single “silver bullet”

RANS methods will continue to see broad application, and may possibly see improvement via RST methodologies

Hybrid RANS-LES methods show the most promise as a good compromise between accuracy and affordability better theoretical approaches for interface region are needed

LES method development is an active area of research and is progressing significant investment still needed to enable the technology for broad engineering application in 2030

Continued investment needed for the development of a validated, predictive, multi-scale combustion modeling capability (e.g., to optimize the design and operation of evolving fuels for advanced engines)

11

Wall-modeled LES (WMLES) cost estimates

Using explicit, 2nd order accurate finite volume/difference

Unit aspect ratio wing, Mach 0.2 flow

Comparison to current HPC #1 system: Tianhe-2

55 PFLOP/s theoretical peak; 34 PFLOP/s on Linpack benchmark

WMLES Re=1e6 feasible today on leadership class machines

2030 HPC system estimate

30 ExaFLOP/s theoretical peak

WMLES Re=1e8 feasible on 2030 HPC

Comments:

These are capability computations (maxing out leadership HPC)

Simple geometry (unit aspect ratio; isolated, clean wing; etc)

Algorithmic advances critical for grand challenge problems

Case Study: LES Cost Estimates

12

24 hour turn-

around time

Findings CONTINUED

4. Mesh generation and adaptivity continue to be significant bottlenecks in the CFD workflow, and very little government investment has been targeted in these areas.

Streamlined and robust geometry (e.g., CAD) access, interfaces, and integration into CFD processes is lacking

Large-scale, automated, parallel mesh generation is needed as the size and complexity of CFD simulations increases goal is to make grid generation invisible to the CFD analysis process

Robust and optimal mesh adaptation methods need to become the norm

Curved mesh element generation for higher-order discretizations is needed

Consider newer strategies like cut cells, strand grids, “meshless”

13

Findings CONTINUED

5. Revolutionary algorithmic improvements will be required to enable future advances in simulation capability.

Robust convergence behavior for complex geometries and flowfields is lacking need automated, mesh-tolerant, monotone positivity-preserving and entropy-preserving schemes

Improved scalability of CFD algorithms on current and emerging HPC hardware is needed develop “optimal” solvers, improve discretizations (e.g., higher-order)

14

Robust uncertainty quantification methods are needed

– Extension of output-based error techniques and grid adaptation is not robust for 3D viscous flows about complex geometries

– Current output-based error estimates can be significantly inaccurate and/or produce unbounded errors

– Error methods for parametric variability (e.g., boundary and initial conditions)

Basic research in optimal scalable solvers required to Advance numerical efficiency of CFD solvers (basic R&D)*

Capitalize on emerging HPC hardware (HPC access)*

Asymptotic properties of optimal solvers: Gains increase with problem size

Gains increase with increasing HPC parallelism

Orders of magnitude possible for large problems

Many agencies have focused scalable solver projects e.g. DoE ASCR program, CERFACS

Multigrid used as example Initially developed with NASA funding (Brandt & South, Jameson)

– NASA codes utilize older NASA developed multigrid (limitations)

Considerable on-going development outside of NASA – LLNL AMG solvers demonstrated on 100,000 cores, 1012 DOFs

– Not always directly transferrable to aero/CFD problems

Solver development should focus on techniques best suited for important NASA problems Hyperbolic, time-implicit, stiff source terms…

Case Study: Scalable Solvers

15 *Resources required

Findings CONTINUED

6. Managing the vast amounts of data generated by current and future large-scale simulations will continue to be problematic and will become increasingly complex due to changing HPC hardware.

Tools to effectively visualize, manage, and store single, very-large, high-fidelity simulations must improve development of innovative visualization and presentation techniques

Need for the processing (collection, synthesis, and interrogation) of thousands of CFD simulations in real-time ROM, “meta-models”, etc.

Development of methods to merge CFD data with other aerodynamic source data (e.g., wind tunnel, flight data) and multi-disciplinary simulation data to create an integrated database (with confidence level) is critical

16

Findings CONTINUED

7. In order to enable increasingly multidisciplinary simulations, for both analysis and design optimization purposes, several advances are required:

Individual component CFD solver robustness and automation will be required.

Development of standards for coupling of CFD to high-fidelity simulations of other disciplines

Emphasis on the Science of MDAO and the development of stable, accurate and conservatives techniques for information transfer

Techniques for computing sensitivity information and propagating uncertainties in the context of high-fidelity MDAO problems

17

18

Vision of CFD in 2030

Emphasis on physics-based, predictive modeling Transition, turbulence, separation, chemically-reacting flows, radiation,

heat transfer, and constitutive models, among others.

Management of errors and uncertainties From physical modeling, mesh and discretization inadequacies, natural variability (aleatory), lack of knowledge in the parameters of a particular fluid flow problem (epistemic), etc.

A much higher degree of automation in all steps of the analysis process Geometry creation, mesh generation and adaptation, large databases of simulation results, extraction and understanding of the vast amounts of information generated with minimal user intervention.

Ability to effectively utilize massively parallel, heterogeneous, and fault-tolerant HPC architectures that will be available in the 2030 time frame Multiple memory hierarchies, latencies, bandwidths, etc.

Flexible use of HPC systems Capability- and capacity-computing tasks in both industrial and research environments.

Seamless integration with multi-disciplinary analyses High fidelity CFD tools, interfaces, coupling approaches, etc.

19

Grand Challenge Problems

Represent critical step changes in engineering design capability

May not be routinely achievable by 2030

Representative of key elements of major NASA missions

1. Large Eddy Simulation (LES) of a powered aircraft configuration across the full flight envelope

2. Off-design turbofan engine transient simulation

3. Multi-Disciplinary Analysis and Optimization (MDAO) of a highly-flexible advanced aircraft configuration

4. Probabilistic analysis of a powered space access configuration

Assess the ability to use CFD over the entire flight envelope, including dynamic maneuvers

Assess the ability of CFD to accurately predict separated turbulent flows Monitor increasing LES region for hybrid

RANS-LES simulations

Evaluate success of WMLES

Determine future feasibility of WRLES

Assess the ability to model or simulate transition effects

Project future reductions in wind tunnel testing

20

LES of a Powered Aircraft Configuration Across the Full Flight Envelope

Measure progress towards virtual engine testing and off-design characterization

Assess the ability to accurately predict: Separated flows

Secondary flows

Conjugate heat transfer

Rotating components, periodic behavior

Potential to demonstrate industrial use of WRLES for lower Re regions

Assess progress in combustion modeling and prediction abilities

21

Off-Design Turbofan Engine Transient Simulation

Ultimate utility of CFD for aerospace engineering is as component for MDAO

Future vehicle configurations to be highly flexible

Assess progress in analyzing the important multidisciplinary problems: Science of coupling Time dependent

Aero-structural

Aero-servo-elastic

Aerothermoelastic

Assess multidisciplinary optimization capabilities Availability of sensitivties

Performance of optimizations

Optimization under uncertainties

22

MDAO of a Highly-Flexible Advanced Aircraft Configuration

Opening up new frontiers in space vehicle design hinges on development of more capable high-fidelity simulations

Assess specific relevant capabilities Separated turbulent flows

High-speed/hypersonic flows

Aero-plume interactions

Aerothermal predictions

Emphasis on reducing risk through uncertainty quantification techniques Unique configurations

Limited experience base

Difficult conditions for ground-based testing

23

Probabilistic Analysis of a Space Access Vehicle

24

Technology Roadmap

Visualization

Unsteady, complex geometry, separated flow at

flight Reynolds number (e.g., high lift)

2030202520202015

HPCCFD on Massively Parallel Systems

CFD on Revolutionary Systems

(Quantum, Bio, etc.)

TRL LOW

MEDIUM

HIGH

PETASCALE

Demonstrate implementation of CFD

algorithms for extreme parallelism in

NASA CFD codes (e.g., FUN3D)

EXASCALE

Technology Milestone

Demonstrate efficiently scaled

CFD simulation capability on an

exascale system

30 exaFLOPS, unsteady,

maneuvering flight, full engine

simulation (with combustion)

Physical Modeling

RANS

Hybrid RANS/LES

LES

Improved RST models

in CFD codes

Technology Demonstration

Algorithms

Convergence/Robustness

Uncertainty Quantification (UQ)

Production scalable

entropy-stable solvers

Characterization of UQ in aerospace

Highly accurate RST models for flow separation

Large scale stochastic capabilities in CFD

Knowledge ExtractionOn demand analysis/visualization of a

10B point unsteady CFD simulation

MDAO

Define standard for coupling

to other disciplines

High fidelity coupling

techniques/frameworks

Incorporation of UQ for MDAO

UQ-Enabled MDAO

Integrated transition

prediction

Decision Gate

YES

NO

NO

Scalable optimal solvers

YES

NODemonstrate solution of a

representative model problem

Robust CFD for

complex MDAs

Automated robust solvers

Reliable error estimates in CFD codes

MDAO simulation of an entire

aircraft (e.g., aero-acoustics)

On demand analysis/visualization of a

100B point unsteady CFD simulation

Creation of real-time multi-fidelity database: 1000 unsteady CFD

simulations plus test data with complete UQ of all data sources

WMLES/WRLES for complex 3D flows at appropriate Re

Integrated Databases

Simplified data

representation

Geometry and Grid

Generation

Fixed Grid

Adaptive Grid

Tighter CAD couplingLarge scale parallel

mesh generationAutomated in-situ mesh

with adaptive control

Production AMR in CFD codes

Uncertainty propagation

capabilities in CFD

Grid convergence for a

complete configuration

Multi-regime

turbulence-chemistry

interaction model

Chemical kinetics

in LESChemical kinetics

calculation speedupCombustion

Unsteady, 3D geometry, separated flow

(e.g., rotating turbomachinery with reactions)

Recommendations

1. NASA should develop, fund and sustain a base research and technology (R/T) development program for simulation-based analysis and design technologies.

25

Required to fulfill technology development plan and address Grand Challenge problems

RCA program to coordinate ALL key CFD technologies, including combustion and MDAO structured around six technology areas

Success will require collaboration with experts in mathematics, computer science, computational geometry, and other aerospace disciplines

AERONAUTICAL SCIENCES PROJECT

REVOLUTIONARY COMPUTATIONAL

AEROSCIENCES (RCA)

HPCPhysical Modeling:

TurbulenceTransitionCombustion

Numerical AlgorithmsGeometry/GridKnowledge ManagementMDAO (Interfaces/coupling)

STRUCTURES ANDMATERIALS

COMBUSTIONCONTROLSINNOVATIVE

MEASUREMENTSMDAO

New thrust

Shared investment and technology collaboration

AERONAUTICS RESEARCH MISSION DIRECTORATE (ARMD)

SCIENCE MISSION

DIRECTORATE

(SMD)

HUMAN

EXPLORATION AND

OPERATIONS

DIRECTORATE

(HEO)

FUNDAMENTAL AERONAUTICS PROGRAM

Close coordination

Recommendations CONTINUED

26

HPC

1. Increasing access to

leading-edge HPC hardware

2. Porting of current and future

codes to leading-edge HPC

3. Radical emerging HPC

technologies

PHYSICAL MODELING

1. RANS turbulence model

2. Hybrid RANS-LES modeling

a. Improved RANS

component

b. Seamless interface

3. LES (wall-modeled and wall-

resolved)

4. Transition

5. Combustion

NUMERICAL ALGORITHMS

1. Advances in current algorithms

for HPC

2. Discretizations

a. Higher-order methods

b. Low dissipation/dispersion

schemes

c. Novel foundational

approaches

3. Solvers

a. Linear and non-linear

scalable solvers

b. Enhancements for MDAO

and UQ

4. UQ

a. Define aerospace

uncertainties

b. Leverage known techniques

c. Improved error estimation

techniques

d. Statistical approaches

GEOMETRY AND GRID

GENERATION

1. CAD access and interfaces

2. Large scale parallel mesh

generation

3. Adaptive mesh refinement

KNOWLEDGE EXTRACTION

1. Visualization

2. Data-base management

3. Variable fidelity models

MDAO

1. Interfaces and standards

2. Accurate and stable

coupling techniques

3. UQ support and

sensitivities (system-level)


2. NASA should develop and maintain an integrated simulation and software development infrastructure to enable rapid CFD technology maturation.

Maintain a world-class in-house simulation capability

– Superior to capabilities within academia and industry

– Cover all key vehicle classes (e.g., external aero, turbomachinery, science and space)

– Critical for understanding principal technical issues, driving development of new techniques, and demonstrating capabilities

Streamline and improve software development processes

– Create formal software development, maintenance, and testing environment to build new and/or enhance existing NASA CFD tools and processes streamlines efforts, synergizes development teams, reduces costs

Emphasize CFD standards and interfaces

– Emphasize common ways of exchanging data (e.g., geometry, grid, MDAO) streamlines and optimizes development efforts

– Must identify ALL key stakeholders, develop consensus, and advocate for standard approaches

27


3. NASA should make available and utilize HPC systems for large-scale CFD development and testing.

Provide access to large-scale computing for both throughput (capacity) to support on-going programs, but also development (capability) to directly support technology demonstrations and progress towards Grand Challenge problems.

– Mitigates stagnating simulation capabilities

– Drives progress by testing new algorithms at scale

– Develops a more experienced and effective user base both within and outside NASA

Leverage existing internal, as well as other national HPC resources

Provide access to novel HPC platforms (e.g., D-Wave) to accelerate technology development

28


4. NASA should lead efforts to develop and execute integrated experimental testing and computational validation campaigns.

High quality experimental test data for both fundamental, building-block and complex, realistic configurations, coupled with careful computational assessment and validation, is needed to advance CFD towards the Vision 2030 goals

NASA is uniquely positioned to provide key efforts in this area due to the availability of experimental test facilities and experience, as well as key expertise to benchmark CFD capabilities

29


5. NASA should develop, foster, and leverage improved collaborations with key research partners and industrial stakeholders across disciplines within the broader scientific and engineering communities.

Leverage other government agencies and stakeholders (US and foreign) outside of the aerospace field collaborate with DoE, DoD, NSF, NIST, etc.

30

Embrace and establish sponsored research institutes acts as a centralized focal point for the development of cross-cutting disciplines, engages the broader scientific community, and executes a long-term research strategy

Actively pursue collaborations with both academia AND industry Enables accelerated efforts for CFD advances in full flight envelope, reduction of physical testing, and certification by analysis

Emphasize basic funding in applied math and computer science Advanced developments in CFD will require breakthroughs in numerical algorithms and efficient solution techniques for emerging HPC systems

Research institutes provide an effective mechanism for: Engaging broader scientific community Centralizing multidisciplinary & cross-cutting technology R&D Enable long-term focus while aligning with agency missions

Self supporting institutes deemed unfeasible Compromises long term focus Poorly aligned with NASA missions

Examples of aerospace computational science institutes CERFACS

– Multidiscplinary: aero, climate, computer science – Includes public and private funding (6 shareholders)

C2A2S2E – Highly focused on external aeronautics – Includes public and private funding

ICASE – Highly successful – Long term focus required core funding and independent agenda

Institute structure and strategy more important than funding levels Must be considered carefully

Case Study: Sponsored Research Institutes

31


6. NASA should attract world-class engineers and scientists.

Success in achieving the Vision 2030 CFD capabilities is highly dependent on obtaining, training, and nurturing a highly educated and effective workforce need for researchers, developers, and practitioners from diverse backgrounds

How?

– Expand fellowship programs in key computational areas

– Encourage and fund long-term visiting research programs

– Provide visiting students with meaningful opportunities to make progress towards the solution of Grand Challenge problems

32

33

Answers to the Key NASA Questions

1. What hardware requirements and software attributes will characterize an advanced “Vision 2030” CFD code that is used to routinely compute complex turbulent flows, involving flow separation, at subsonic, supersonic, and hypersonic speeds, and exhibits greater robustness than current technology?

HPC systems that provide a dramatic increase in computational power, are fault tolerant, and are available and affordable for the routine, production design and analysis of aerospace vehicles and systems

CFD flow solvers that:

– run efficiently on future exascale systems,

– predict vehicle/system performance with certifiable accuracy and full quantification of errors,

– utilize optimal solution schemes and intuitive, parameter-free interfaces,

– fully incorporate mesh generation and adaptation as part of the solution,

– employ advanced turbulence and transition models that predict all types of flow separation,

– are seamlessly integrated into visualization and data mining techniques, and

– are effectively coupled with other disciplines in computational mechanics

34

Answers to the Key NASA Questions CONTINUED

2. How many orders of magnitude faster (time-to-solution), as compared with current capabilities, will 2030 CFD technology be? Expect 1000-10000x increase in computational power which will enable:

– 1011 grid point or equivalent high-order resolution full-configuration steady-state RANS simulations in minutes on commodity hardware, or 1012 grid point simulation in seconds on leadership hardware

– Equivalent highly resolved steady-RANS based optimization in minutes to hours

– Highly resolved time-dependent production RANS simulations (e.g., full rotorcraft simulations at high spatial and temporal resolution (1/10th degree time step, hundred revs) in hours on commodity hardware)

– Increased emphasis on including more complex physics into the simulations

– Generation of a complete, CFD-based aerodynamic database (~10,000 data points) in under one day using steady RANS (for all flight regimes)

– Tightly coupled multidisciplinary simulations (e.g. aeroelastic, aerothermoelastic, aeroservoelastic, maneuvering aircraft/digital flight) and optimizations at varying levels of fidelity overnight on commodity hardware and in hours on leadership hardware

35


3. What will be the fundamental elements of 2030 turbulence models, and for what classes of problems will they be reliable?

Hybrid RANS-LES methods will continue to improve – RANS will cover the area where the boundary layer (BL) is too thin for LES

(and the area will shrink as computing power rises) – Simulations at low Reynolds numbers may be treated with LES only – All high Reynolds number LES will be Wall-Modeled LES

RANS models will remain very empirical – Reynolds-Stress Transport models will improve… – …but may or may not retire “few-equation” models, especially non-linear versions

Transition prediction will become much more routine – en methods and those based on empirical transport equations will co-exist – The primary challenge is integration of en methods with flow solvers

The accuracy of aerodynamics CFD will be competitive with wind tunnels – Fair or better prediction of BL transition and separation (RANS + LES) – Fair or better predictions after massive separation (LES) – Affordable for attached flows; expensive for wide range of scales (e.g., AFC) – Full Reynolds number and Mach number with access to flow field data leading to

better understanding of flow physics through advanced visualization

CFD combustion predictions will be less competitive (than aero) – Turbulent combustion models will remain challenged by modeling at the subgrid

scale

36



Effective utilization of HPC – Need for scalable solver/algorithm/pre-post-processing development – Software issues/programming paradigm – Lack of access to HPC for scalable software development and testing

Physical modeling – Turbulent separated flow:

o RANS models stagnant and inadequate o Hybrid RANS-LES rely on RANS and require better RANS-LES

interface o Sheer cost of LES with current discretizations/solvers

– Transition modeling – Combustion

37



Autonomous and reliable CFD – Cumbersome CAD/mesh/adaptivity processes – Lack of optimal and scalable solvers (impediment grows with problem

size) – Unreliable error estimates – Uncertainty quantification implemented as after-thought or not at all

Knowledge extraction and visualization – Effective analysis of large simulation (e.g. parallel co-visualization) – Real time processing of entire simulation databases – Merging of variable fidelity simulation data and experimental data

Multi-disciplinary/multi-physics simulations and frameworks – Robustness and automation of CFD within MDAO – Science of multidisciplinary coupling at high fidelity – Fully coupled sensitivities – Frameworks/standards

38


5. What high-risk/high-yield obstacles remain as enduring and daunting challenges that should be the focus of long-range efforts by NASA?

Daunting challenge: Develop reliable physics-based high-fidelity MDAO tools for aerospace

vehicle analysis, design, and reduce risk through quantifiable uncertainties o For aeronautics and space vehicles

High-Risk/High-Yield: Our listed obstacles show that this will not be achieved by investment

in one or two specific technologies (e.g. current LES approaches remain out of reach due to cost considerations)

Requires investment in broad range of synergistic technologies to bring step changes in simulation capability to fruition o Foundational (applied math, computer science, physical modeling) o Combine algorithmic advances with HPC advances

Organizational, cultural, and collaborative challenges are just as important as technical challenges and must be addressed

Backup

39

Thoughts on Revolutionary Approaches

Advances in CFD between now and 2030 will likely include both evolutionary improvements to current tools, methods, and processes, and revolutionary, “paradigm-shifting” advances in technology.

Although revolutionary breakthroughs cannot be predicted, emerging technologies that promise significant advances in capability, such as quantum or bio-computing, must be properly followed and nurtured.

Evolutionary improvements in CFD technology (e.g., upwind shock capturing algorithms) do lead to revolutionary advances in simulation capability (e.g., routine prediction of cruise drag).

40

Unstructured mesh technology development:

• First demonstration of full aircraft CFD using unstructured mesh (full potential) : Dassault Aviation 1982

• First demonstration of full aircraft CFD using unstructured mesh (Euler equations): Jameson 1985

• NASA unstructured mesh development: initiated in 1990's (FUN3D)

• Unstructured mesh RANS solvers dominant approach (e.g. DPW series by late 2000s)

Objectives

Collect critical feedback from the engineering and scientific communities on key topics related to CFD

Develop definitive views on CFD requirements, shortcomings, and impediments

Identify key discussion topics for CFD workshop

Conducted during February 2013

Hosted by SurveyMonkey

175 total participants

22 total questions (14 technical)

Excellent feedback (especially the written responses)

41

CFD Survey

1. CFD requirements captured well by Vision 2030 CFD team

2. 68% would use increases in HPC for MDA/O and multi-physics analysis

3. Most important enhancement to CFD: Increased accuracy of flow physics modeling

4. Critical impediments: Lack of validation datasets, academia/govt/industry coordination, CFD best practices

5. Primary focus of turbulence model development: Enhancement to hybrid RANS/LES models

6. User aspects: More automation, faster visualization, need for hard error estimates, more feedback (solver, trouble areas)

7. MDA/O: Most critical discipline to couple – Structures

8. MDA/O: Key issue – Integration

42

CFD Survey – What We Learned

CFD Survey – What We Learned

9. Revolutionary capabilities: Certification, maneuvering flight, full flight envelope + aeroelastics + complete engine effects

10. CFD software development: Better interaction between math/CS/engineering, better software engineering practices

11. HPC: Speed-up likely through parallelism, Quantum computing unlikely to be ready by 2030

12. Long term development plan: Improvements to physical modeling and numerical algorithms (44%), streamlining and error-proofing of CFD processes (25%), enhancements to parallel computing efficiency (19%)

13. Use and impact: CFD will be easier for non-experts, CFD will be used earlier in the design process, CFD will reduce/replace physical testing

14. Development and validation of large-scale CFD simulations

43

CFD Workshop

Develop consensus view of key requirements and impediments

Held 1-2 May 2013 at NIA in Hampton, VA

50+ attendees (16% academia, 46% government, 32% industry, 6% CFD software vendors)

7 sessions each covering a specific discussion topic:

–Revolutionary Thoughts –HPC Trends, Software Development –Turbulence Modeling, Separated Flows, etc. –Numerics, Error Quantification, Solution Quality –Geometry, Grid, CFD Process –Complex Flow Physics –MDA/O, Multi-Physics

44

Unsteady, turbulence/transition/separated flows Numerical efficiency and accuracy

Autonomous and reliable CFD Geometry Grid generation, adaptation, convergence UQ: Modeling, Numerical, Natural Variability

Knowledge extraction Data mining, data fusion (merging CFD/test, etc.), “Big Data”

HPC challenges

Interface/frameworks for MDA/O, multi-physics

CFD Validation

Coordination/Collaboration

Education of the CFD Workforce

45

CFD Workshop – Key Issues

“Surprise”

46

Team Structure

Boeing

Research & Technology

Greg Hyslop, President

HPC NCSA

William Gropp

Focal

Propulsion Pratt & Whitney

Elizabeth Lurie

Focal

Airframe Boeing

Jeffrey Slotnick Principal Investigator

Flight Sciences

Technology

Mark Anderson, Director

Program Manager

Abdi Khodadoust

Academia Research Team

David Darmofal, MIT

Focal

• J. Alonso, Stanford

• D. Mavriplis, Wyoming

• D. Dominik, Boeing

• P. Fussell, Boeing

• J. Gablonski, Boeing

• R. Bush, P&W

• J. Croxford, Boeing

• M. Lamoureux, P&W

• W. Lord, P&W

• R. Shaw, Boeing

• N. Somanath, P&W

• T. Antani, Boeing

• K. Bowcutt, Boeing

• M. Mani, Boeing

• R. Narducci, Boeing

• P. Spalart, Boeing

• J. Vassberg, Boeing

• V. Venkatakrishnan, Boeing

NASA COTR

William Kleb

Boeing Business Team

Kurt Bratten

NASA and

Government Lab

Researchers

• Sandia

• AFRL

• …

Technical Advisory Team

Mark Anderson, Leader

47

Statement of Work 3.2 Initial Vision of 2030 CFD Code

3.2.1 Initial Vision Development

Collect and establish requirements (Deliverable 2) Identify shortcomings and gaps in current CFD technology

Consider interplay of 2030 CFD Code with multi-physics analysis capabilities, quantification of computational uncertainty, and emerging High Performance Computing (HPC) infrastructure

Identify impediments to achieving the 2030 Vision CFD Code

3.2.2 Survey of CFD Community Develop and conduct comprehensive technical survey to pulse CFD and

scientific communities on CFD requirements, gaps, limitations, and impediments

Develop list of CFD limitations, gaps, and impediments (Deliverable 3)

‒ NASA Centers ‒ Government Labs ‒ Aerospace Industry

‒ Academic Institutions ‒ HPC Community ‒ CFD Vendors

48

Statement of Work 3.3 Refined Vision of 2030 CFD Code

3.3.1 Collect Survey Feedback (Deliverable 4)

Create a comprehensive summary of CFD community survey results

Obtain feedback from CFD community on survey results

3.3.2 Vision 2030 CFD Code Workshop

Plan and host a 2-3 day joint workshop between Boeing team

members, NASA FAP and CFD leaders, and the broader CFD

community – Establish a definitive view on requirements, priorities, roadblocks, and

impediments

– Develop realistic technical approaches and generate new ideas in improving CFD capability

– Assemble consensus and actionable plans for maturing CFD technologies (using the TRL scale), including identifying validation activities and computational resources

– Develop initial timeline for required research

49

Statement of Work 3.4 Mid-Term Review

Present a finalized vision of the 2030 CFD Code

Present an outline of the initial research plan and technology development roadmap

Develop and deliver mid-term review presentation package (Deliverable 5)

50

Statement of Work 3.5 Vision 2030 CFD Code Integrated Plan and Roadmap

3.5.1 Develop Detailed Research Plan and Roadmap Technology trends

Anticipated technology challenges

Success metrics and appropriate exit criteria

Estimated 2030 code performance

3.5.2 Develop System-Level View of Architecture Numerical algorithms and solution technology

Software design

Acknowledge and consider likely interconnectivity with future multi-disciplinary design/optimization and analysis systems

3.5.3 Answers to the Key NASA Questions

51

Statement of Work 3.6 Final Review

Present a complete summary of the Vision 2030 CFD Code including comprehensive research plans and technology roadmap

Develop and deliver final review presentation package (Deliverable 6)

52

Statement of Work 3.7 Final Report

Provide a written report for the complete summary of the Vision 2030 CFD Code including comprehensive research plans and technology roadmap (Deliverable 7)

53

Deliverables

# DELIVERABLE TASK

REF

FORMAT DATE

1 Kickoff meeting presentation

package

3.1 Presentation

MS Powerpoint 26 October 2012

2 Initial 2030 CFD Code

Requirements

3.2.1 Informal Report

MS Word 28 November 2012

3 CFD limitations, gaps, and

impediments

3.2.2 Informal Report

MS Word 21 December 2012

4 Survey Input Summary 3.3.1 Informal Report

MS Word 28 February 2013

5 Mid-term Technical Review

Presentation Package

3.4 Presentation

MS Powerpoint 27 March 2013

6 Final Technical Review

Presentation Package

3.6 Presentation

MS Powerpoint 28 August 2013

7(a)

7(b)

Final written report (draft)

Final written report

3.7 NASA Contractor

Report Format 28 August 2013

27 September 2012

8 Monthly Project Telecon

minutes

5.2 Informal Report

MS Word 5 working days after

meeting

9 Monthly Progress Reports 5.3 Informal Report

MS Word 15th calendar day of

following month

54

Meeting & Teleconference Schedule

REF MEETING DATE DURATION LOCATION

3.1 Kick-Off Meeting 31 October 2012 1 Day NASA Langley

4.1 Monthly Project

Meeting Telecons

3rd Thursday of each

month (8am-9am PST)

1 hr Telephone

WEBEX

3.3.2 2030 Vision CFD Code

Workshop

5-7 March 2013 3 Days TBD

3.4 Mid-Term Technical

Review

27-28 March 2013 2 Days NASA Langley

3.6 Final Technical

Review Meeting

24-25 September 2013 2 Days NASA Langley

PROPOSED

55

Technical Approach

2030 Today

Determine impediments

and roadblocks

Detailed Research Plan and Technology Roadmap

HPC Establish expected

computing performance

and likely outlook for CFD

technology application

Academics Gov’t Labs Industry Labs Identify current CFD technology

gaps and shortcomings

Industry Define anticipated aerodynamic

and performance

requirements for competitive

aerospace vehicle and systems

NASA Requirements developed

for N+2/N+3 systems,

including noise,

emissions, fuel

consumption targets, etc.

DoD

nasa vision 2030 cfd code · 2018-01-24 · nasa contracting officer representative – bil kleb...

Documents